JPH06103001A - Magnetic disk sub system - Google Patents

Abstract

PURPOSE:To provide the compact magnetic disk system which enables the multiple control of plural disk devices, accelerates write/read speed, improves data reliability and reduces the rotation waiting of a disk controller. CONSTITUTION:The multiple control is enabled by dual ports [disk side ports 8 and 8 and interfaces (a) and (b)] provided at a disk controller 1, the data reliability is improved by composing a cache memory 2 of a non-volatile memory or a battery backup memory, and high-speed write/read is enabled by forming the interfaces (a) and (b) as the parallel transfer interfaces having an 8-bit width or 16-bit width. Further, it is not necessary for the disk controller to wait for rotation by reporting only the end of search following the seek end of disk from the interfaces (a) and (b) to the disk controller 1 without reporting the seek end of the disk. Then, a host side port 6 can be made dual as well.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file device used in a computer system, and more particularly to a small magnetic disk subsystem which is highly reliable and capable of high speed recording / reproducing processing.

[0002]

2. Description of the Related Art Conventionally, for example, JP-A-59-135563.
In the magnetic disk subsystem, as described in Japanese Patent Publication (Kokai), a plurality of magnetic disk devices that are connected to a higher-level device via a disk controller are configured to be used as random access files. Are known. This configuration is shown in FIG.

In FIG. 2, the disk controller 21 is
Upper device (not shown) via upper interface 24
Disk interface 27
It is connected to a plurality of magnetic disk devices 25 via. In the disk controller 21, a CPU 23 that controls the entire controller and a cache memory 22 that temporarily holds data are provided. Reference numeral 26 is a host interface circuit (host side port), 27 is a drive interface, 28 is a drive interface circuit (drive side port), 29 is an internal bus, and 30 is an internal control signal.

When executing the READ command of the host device, the control device 21 performs seek control of the head of the sector of the target disk device 25 in which the target data is recorded and rotation waiting control up to the target sector. After that, the data is read from the disk device 25 into the cache, and the data is transferred to the host device. In this way, RE of normal data
When performing AD, it takes time to access the disk,
As a solution, the data is held in the cache, and the next time a READ command instruction arrives from the host device to the same address, without accessing the disk,
The host response time has been accelerated by transferring the data held in the cache memory. However, in the conventional technique, the W from the host device is always used at the time of the WRITE command.
After writing the RITE data to the disk device,
It reported the completion to the host device. When this is written to the cache memory and the end is reported, then a power failure,
This is because data cannot be guaranteed when the cache memory is destroyed due to a momentary interruption or the like. For this reason, in the conventional technique, since the disk access is always performed when the WRITE command is executed, there is a problem that seek or rotation wait occurs and the response speed is poor.

[0005]

In the above-mentioned prior art, a plurality of magnetic disk devices are connected, but since the control of only a plurality of magnetic disk devices is taken into consideration, the performance, function and usability of the entire system have not been taken into consideration. . That is, in the above conventional technique,
(1) Since the disk control device and the plurality of magnetic disk devices are connected by only one drive-side port or drive interface, multiple magnetic disk devices cannot be simultaneously multiplexed and controlled (2 ) The drive interface is a 1-bit serial transmission type, so the data transfer rate is low, (3) the cache memory has a volatile scale, so the data reliability is low, and (4) as described above. In addition, when the write command is executed due to the volatile cache memory, the disk is accessed and the data is written to the disk before the completion of writing is reported to the host device. However, there was a problem that the response speed was slow.

Therefore, an object of the present invention is to solve the above-mentioned problems of the prior art, to enable multiple control of a plurality of magnetic disk devices, to achieve high data write / read processing speed, high data reliability, and arbitrary The purpose of the present invention is to provide a magnetic disk system capable of writing to a disk at the time of, and having less waiting for rotation of a host device or a disk controller.

[0007]

In order to achieve the above object, the present invention provides a magnetic disk device and a disk which receives a recording or reproducing command from a host device and controls the recording or reproducing of data by the magnetic disk device. A magnetic disk subsystem having a control device, wherein the disk control device is provided with a data cache memory and a drive-side port, and the drive-side port and the magnetic disk device are connected via a drive interface, At least two sets of drive-side ports and drive interfaces are provided (dual port configuration), and the drive interface is configured by a parallel data transfer interface.

The cache memory is composed of a non-volatile memory or a memory backed up by a battery.

Further, the disk control device may be provided with at least two sets of host interface ports for host devices.

Further, the drive interface does not report the end to the disk control device when the seek of the magnetic disk device is completed, and when the target rotational position of the magnetic disk device is detected (when the search is completed). , Configure it to report its detection.

[0011]

The operation based on the above configuration will be described.

According to the present invention, since the interface with the magnetic disk device has a dual port configuration, it is possible to perform multiplex processing on a plurality of magnetic disk devices. Further, by configuring this interface with a parallel data transfer interface having an 8-bit data width or a 16-bit data width, for example, recording / reproducing processing for the disk and data transfer can be performed at high speed. As a result, a high performance and highly reliable magnetic disk system can be realized.

Further, as the cache memory, a nonvolatile memory or a low power consumption R backed up by a battery
By using a memory such as AM, write data is guaranteed even when the power is turned off, and a write completion report is sent to the host device when writing to the cache memory, and the cache data is written to the disk at any time. It can be so.

Further, since the drive interface does not send the seek end report to the disk control device but sends the search end report, it is not necessary to wait for the disk control device to rotate in the device above, and the processing efficiency of the system is improved. Can be improved.

[0015]

Embodiments of the present invention will be described in detail below with reference to the drawings.

First, FIG. 1 shows the configuration of a disk control system according to a first embodiment of the present invention. The disk controller 1
It is connected to a host device (not shown) via the host interface 4. The disk device 5 is connected via a disk interface (drive interface) 7 having a dual port. In the disk controller 1, the CPU 3 that controls the entire controller 1
And a non-volatile (NVS) cache memory 2 for temporarily holding data from the host or disk. In addition, 6 is a host interface port (host side port), 8 is a drive interface port (drive side port), 9 is an internal data bus, 10
Is an internal control signal, and a and b are drive interfaces.

A READ command from the host is interpreted by the CPU 3 of the control device 1 via the interface 4. Then, a seek command to the target address is issued to the disk device 5 via the disk interface 7, and after waiting for rotation up to the sector, the data is read to the cache memory 2. Then, the data is transferred to the host.
Since the data is held in the cache 2, when the READ command of the next host has the same address, only the data transfer from the cache is performed and the command is terminated without accessing the disk. As a result, the response time of the host is shortened. Further, according to the present embodiment, the READ command shown before is also applied to the WRITE command which has always required the disk access.
Similar to the command, since the completion of the WRITE command is reported to the host before the disk is accessed, the response time to the host can be shortened. It is shown below.

In the present embodiment, in response to the WRITE command from the host, the end can be reported to the host when the data from the host is written in the cache memory. this is,
Since the cache memory is NVS, the data in the cache memory can be retained even in the event of a power failure or momentary interruption.
This is because it can be guaranteed. Then, the time-consuming seek to the disk, the waiting for rotation, and the WRITE operation are executed in the idle time when the command of the host is not issued and at the timing asynchronous with the host. As a result, the response time to the host can be shortened even during the WRITE operation,
Moreover, data can be guaranteed and reliability can be secured even in the unlikely event of a power failure or momentary interruption. Further, in this embodiment, since two sets of the disk interface port 8 to the disk interface 7 are provided, for example, when data is being transferred by the interface a to the disk device "1", the interface b is used. Thus, multiple control such as issuing a seek command to the disk device "2" is possible. Therefore, the data from one disc is
W the data on the cache to the other disk during AD
Efficient control such as RITE can be performed and a high-performance and high-performance disk system can be realized.

Also, the drive interface 7 itself has conventionally been ESDI (Enhanced SmallDrive Interface).
Was used. Since this ESDI has one data transfer line, it is a serial transfer of 1 bit at a time, and there is a limitation that there is a limit in high speed data transfer. On the other hand, in the present embodiment, parallel transfer having a plurality of data transfer lines, usually having an 8-bit width, is possible. As a result, high-speed data transfer of the drive interface 7, which has a limit in the past, becomes possible, and higher performance can be realized. This interface is assumed here to be Modified-ESDI.
I am calling.

By using the above non-volatile cache memory, dual-port drive interface, and parallel transfer drive interface 7, a high-performance, high-performance disk system capable of high-speed response can be obtained, and non-volatile. Therefore, it is possible to realize a highly reliable system that can guarantee data even in the event of a power failure.

With respect to the host device as well, two or more sets of host interface ports 6 can be provided to connect to the same or different host devices, and the data processing efficiency can be increased by multiplexing.

FIG. 3 shows a second embodiment of the present invention. The disk controller 1 is provided with a volatile cache memory 2 and a battery 11 for battery backup of the cache memory 2. As a result, even if volatile memory is used, in the unlikely event of a power failure or momentary interruption,
Data can be guaranteed by the battery. In addition, this volatile memory has a low power consumption type R such as DRAM or SRAM.
The use of AM also has the advantage that the data guarantee time can be extended.

The drive interface 7 will be described with reference to FIGS.

FIG. 4 shows a conventional ESDI. Between the disk control device 1 and the disk device 5, a plurality of control signals 7-a, WRITE 7-b for data transfer,
Each one of READ7-c is connected. for that reason,
Data transfer was serial transfer, and there was a limit to high-speed transfer. In contrast, FIG. 5 shows a drive interface according to the present invention. DATABUS7- for data transfer
Since d is used, a multi-bit width such as an 8-bit width or a 16-bit width allows parallel transfer. This allows
Higher-speed data transfer becomes possible.

FIG. 6 is an explanatory view when the drive interface 7 is further provided with a rotation waiting function for the target sector. Normally, in the conventional ESDI, when a seek command is issued from the controller (disk controller) 1 to the disk device, the seek end signal is reset and the seek operation is started, and the disk device switches between seek and head, and moves. The seek end signal is set at the time when the positioning of the head with respect to the target cylinder is completed (point A in the figure). At this time, it is unknown from the controller which rotation position the target sector is, and the controller further
The search waits for the To period (to be exact, the To + T period) until the target sector (its leading position is B) is found by searching D. On the other hand, in the drive interface according to the present invention, no report is made to the controller 1 at the time point A when the seek operation is completed, and the drive interface continues to search the target sector on the disk device side, The end of the search is reported from the disk device side to the controller 1 at time C before the target sector (its head position B) instructed by
It goes without saying that the magnetic disk device 5 has a rotation position detection mechanism that detects that the magnetic disk device 5 has rotated to the target rotation position).

As a result, the controller 1 finds the target sector B immediately at the time point C, so that the time points A to
The period To of C is released from the waiting for rotation, and other control operations can be performed during this period (to be more precise, during the period from issuing the seek command until time C). As a result, the controller does not need to wait for the conventional rotation of To, and high performance is possible.

[0027]

As described in detail above, according to the present invention, since the dual port disk device is connected to the disk control device, during the disk READ of one,
On the other hand, multiple processing such as WRITE to another disk is possible and performance can be improved. Moreover, since the high-speed drive interface is adopted, high-speed command execution and data transfer can be performed, so that the function, performance, and reliability of the entire system can be improved.

Since the non-volatile memory or the battery backup memory is used as the cache memory,
When the host system transfers the WRITE data to the disk system, the completion report can be sent to the host at the time of WRITE to the cache memory, and the disk is not accessed, resulting in a high-speed response system. Moreover, since the battery is backed up, the data in the cache memory can be guaranteed even if the power is cut off.
Since it is a low power consumption RAM, it can be held and guaranteed for a long time.

Further, since the drive interface does not issue the seek end report but the search end report to the magnetic disk control device, there is no need to wait for rotation in a device above the disk control device, and the processing efficiency of the system is improved. To do.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a disk control system according to a first embodiment of this invention.

FIG. 2 is a block diagram of an example of a conventional disk control system.

FIG. 3 is a configuration diagram of a disk control system of a second embodiment of the present invention.

FIG. 4 is a diagram showing a control signal of a drive interface according to the related art.

FIG. 5 is a diagram showing control signals of a drive interface according to an embodiment of the present invention.

FIG. 6 is a timing chart of disk control of the drive interface according to the embodiment of the present invention.

[Explanation of symbols]

1 Disk Controller 2 Cache Memory 3 CPU 4 Host Interface 5 Magnetic Disk Unit 6 Host Interface Port (Host Port) 7 Drive Interface (Disk Interface) 8 Drive Interface Port (Drive Port) 9 Internal Data Bus 10 Internal Control Signal 11 Battery 21 Disk control device 22 Cache memory 23 CPU 24 Host interface 25 Disk device 26 Host interface port (host side port) 27 Drive interface (disk interface) 28 Drive interface port (drive side port) 29 Internal data bus 30 Internal control signal

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Minoru Yoshida Minoru Yoshida 2880 Kozu, Odawara City, Kanagawa Stock Company Hitachi Ltd. Odawara Factory

Claims (4)

[Claims] 1. A magnetic disk device, and a disk control device for controlling recording or reproduction of data by the magnetic disk device in response to a recording or reproduction command from a higher-level device. In a magnetic disk subsystem in which a cache memory and a drive-side port are provided, and the drive-side port and the magnetic disk device are connected via a drive interface, at least two sets of the drive-side port and the drive interface are provided. A magnetic disk subsystem characterized in that the drive interface is configured by a parallel data transfer interface. 2. The magnetic disk subsystem according to claim 1, wherein the cache memory is a non-volatile memory or a memory backed up by a battery. 3. The magnetic disk subsystem according to claim 1, wherein the disk control device is provided with at least two sets of host interface ports for host devices. 4. The drive interface does not report the end to the disk control device when the seek of the magnetic disk device is completed, but reports the detection when the target rotational position of the magnetic disk device is detected. 4. The magnetic disk subsystem according to claim 1, wherein the magnetic disk subsystem is configured as described above.